Centre de documentation
scientifique

Spatial regression graph convolutional neural networks: A deep learning paradigm for spatial multivariate distributions / Di Zhu in Geoinformatica, vol 26 n° 4 (October 2022)  

Public [article]  inGeoinformatica > vol 26 n° 4 (October 2022) . - pp 645 - 676 Titre : Spatial regression graph convolutional neural networks: A deep learning paradigm for spatial multivariate distributions Type de document : Article/Communication Auteurs : Di Zhu, Auteur ; Yu Liu, Auteur ; Xin Yao, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : pp 645 - 676 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géomatique [Termes IGN] analyse multivariée [Termes IGN] analyse spatiale [Termes IGN] apprentissage profond [Termes IGN] distribution spatiale [Termes IGN] échantillonnage [Termes IGN] intelligence artificielle [Termes IGN] régression [Termes IGN] régression géographiquement pondérée [Termes IGN] réseau neuronal convolutif [Termes IGN] réseau neuronal de graphes Mots-clés libres : Geospatial artificial intelligence (GeoAI) Résumé : (auteur) Geospatial artificial intelligence (GeoAI) has emerged as a subfield of GIScience that uses artificial intelligence approaches and machine learning techniques for geographic knowledge discovery. The non-regularity of data structures has recently led to different variants of graph neural networks in the field of computer science, with graph convolutional neural networks being one of the most prominent that operate on non-euclidean structured data where the numbers of nodes connections vary and the nodes are unordered. These networks use graph convolution – commonly known as filters or kernels – in place of general matrix multiplication in at least one of their layers. This paper suggests spatial regression graph convolutional neural networks (SRGCNNs) as a deep learning paradigm that is capable of handling a wide range of geographical tasks where multivariate spatial data needs modeling and prediction. The feasibility of SRGCNNs lies in the feature propagation mechanisms, the spatial locality nature, and a semi-supervised training strategy. In the experiments, this paper demonstrates the operation of SRGCNNs with social media check-in data in Beijing and house price data in San Diego. The results indicate that a well-trained SRGCNN model is capable of learning from samples and performing reasonable predictions for unobserved locations. The paper also presents the effectiveness of incorporating the idea of geographically weighted regression for handling heterogeneity between locations in the model approach. Compared to conventional spatial regression approaches, SRGCNN-based models tend to generate much more accurate and stable results, especially when the sampling ratio is low. This study offers to bridge the methodological gap between graph deep learning and spatial regression analytics. The proposed idea serves as an example to illustrate how spatial analytics can be combined with state-of-the-art deep learning models, and to enlighten future research at the front of GeoAI. Numéro de notice : A2022-865 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE/INFORMATIQUE Nature : Article DOI : 10.1007/s10707-021-00454-x Date de publication en ligne : 02/11/2021 En ligne : https://doi.org/10.1007/s10707-021-00454-x Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=102158 [article]

Semantic enrichment of secondary activities using smart card data and point of interests: a case study in London / Nilufer Sari Aslam in Annals of GIS, vol 27 n° 1 (January 2021)  

Public [article]  inAnnals of GIS > vol 27 n° 1 (January 2021) . - pp 29 - 41 Titre : Semantic enrichment of secondary activities using smart card data and point of interests: a case study in London Type de document : Article/Communication Auteurs : Nilufer Sari Aslam, Auteur ; Di Zhu, Auteur ; Tao Cheng, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : pp 29 - 41 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Analyse spatiale [Termes IGN] carte à puce [Termes IGN] collecte de données [Termes IGN] données socio-économiques [Termes IGN] données spatiotemporelles [Termes IGN] enrichissement sémantique [Termes IGN] loisir [Termes IGN] Londres [Termes IGN] méthode heuristique [Termes IGN] planification urbaine [Termes IGN] point d'intérêt [Termes IGN] transport urbain Résumé : (auteur) The large volume of data automatically collected by smart card fare systems offers a rich source of information regarding daily human activities with a high resolution of spatial and temporal representation. This provides an opportunity for aiding transport planners and policy-makers to plan transport systems and cities more responsively. However, there are currently limitations when it comes to understanding the secondary activities of individual commuters. Accordingly, in this paper, we propose a framework to detect and infer secondary activities from individuals’ daily travel patterns from the smart card data and reduce the use of conventional surveys. First, we proposed a ‘heuristic secondary activity identification algorithm’, which uses commuters’ primary locations (home & work) and the direction (from & to) information to identify secondary activities for individuals. The algorithm provides a high-level classification of the activity types as before-work, midday and after-work activity patterns of individuals. Second, this classification is semantically enriched using Points of Interests to provide meaningful insights into individuals’ travel purposes and mobility in an urban environment. Lastly, using the transit data of London as a case study, the model is compared with a volunteer survey to demonstrate its effectiveness and offering a cost-effective method to travel demand research. Numéro de notice : A2021-319 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article DOI : 10.1080/19475683.2020.1783359 Date de publication en ligne : 01/08/2020 En ligne : https://doi.org/10.1080/19475683.2020.1783359 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97550 [article]